给定Dataframe：

+---+-----------+---------+-------+------------+
| id|      score|tx_amount|isValid|    greeting|
+---+-----------+---------+-------+------------+
|  1|        0.2|    23.78|   true| hello_world|
|  2|        0.6|    12.41|  false|byebye_world|
+---+-----------+---------+-------+------------+

我想将这些列分解成一个名为“col\u value”的行，并对每一行应用逻辑，这样我就得到如下结果：

+---+------------+--------+---------+----------+-------+
| id|   col_value|is_score|is_amount|is_boolean|is_text|
+---+------------+--------+---------+----------+-------+
|  1|         0.2|       Y|        N|         N|      N|
|  1|       23.78|       N|        Y|         N|      N|
|  1|        true|       N|        N|         Y|      N|
|  1| hello_world|       N|        N|         N|      Y|
|  2|         0.6|       Y|        N|         N|      N|
|  2|       12.41|       N|        Y|         N|      N|
|  2|       false|       N|        N|         Y|      N|
|  2|byebye_world|       N|        N|         N|      Y|
+---+------------+--------+---------+----------+-------+

到目前为止，我用 F.arrays_zip spark 2.4的功能：

from pyspark.sql import functions as F
df.withColumn("cols", F.explode(F.arrays_zip(F.array("score", "tx_amount", "isValid", "greeting")))) \
        .select("id", F.col("cols.*")) \
        .withColumnRenamed("0", "col_value")\
        .withColumn("text", (F.regexp_extract(F.col("col_value"),"([A-Za-z]+)",1)))\
        .withColumn("boolean", F.when((F.col("text")=='true')|(F.col("text")=='false'),F.col("text")).otherwise(F.lit("")))\
        .withColumn("text", F.when(F.col("text")==F.col("boolean"), F.lit("")).otherwise(F.col("text")))\
        .withColumn("numeric", F.regexp_extract(F.col("col_value"),"([0-9]+)",1))\
        .withColumn("is_text", F.when(F.col("text")!="", F.lit("Y")).otherwise(F.lit("N")))\
        .withColumn("is_score", F.when(F.col("numeric")<=1, F.lit("Y")).otherwise(F.lit("N")))\
        .withColumn("is_amount", F.when(F.col("numeric")>1, F.lit("Y")).otherwise(F.lit("N")))\
        .withColumn("is_boolean", F.when(F.col("boolean")!="", F.lit("Y")).otherwise(F.lit("N")))\
        .select("id", "col_value","is_score","is_amount","is_boolean","is_text").show()

没有你我怎么办 F.arrays_zip 使用spark 2.3？